Array
(
    [content] => 
    [params] => Array
        (
            [0] => /forum/index.php?threads/why-fabs-are-being-built-in-the-us-and-china.16607/
        )

    [addOns] => Array
        (
            [DL6/MLTP] => 13
            [Hampel/TimeZoneDebug] => 1000070
            [SV/ChangePostDate] => 2010200
            [SemiWiki/Newsletter] => 1000010
            [SemiWiki/WPMenu] => 1000010
            [SemiWiki/XPressExtend] => 1000010
            [ThemeHouse/XLink] => 1000970
            [ThemeHouse/XPress] => 1010570
            [XF] => 2021370
            [XFI] => 1050270
        )

    [wordpress] => /var/www/html
)

Why Fabs are Being Built in the US and China

benb

Well-known member
The semiconductor industry is a model of globalization, with a few countries playing key roles. Japan is a source of advanced materials and automation necessary for fabs to operate, and is perhaps the most important part of this structure. The modern fab depends on robots and materials and Japan has a dominant presence.

This dominance is part technical, part trade. And trade is all about currencies.

USD.KRW bottomed on 11/30/20 at 1085 KRW to the USD. USD.KRW surged to 1347 today, 8/30/22. It has been setting new highs every week for years, and it’s now 24% higher than 11/30/20.
USD.JPY bottomed later than KRW, on 12/28/20, at 103. Since then USD.JPY has been on an epic rise to near new highs on 8/30/22, at 138. It has been less of a consistent rise with some corrections along the way to 34% rise since the end of 2020.
USD.CNH (Offshore Renminbi) has maintained the peg to the USD within a 6.9-7.1 range over the same period.

This brings me to Korea. Korean vendors have certain niches in materials and equipment like Japan that can be exploited when the Won falls.

So US or Chinese fabs purchasing Korean or Japanese goods in Won or Yen have a growing advantage that is leading to more fab construction in these locations. A 24-34% cost reduction for top quality goods is irresistible, both economically (it is worth billions over the life of the fab) and, I think, emotionally (FOMO). Conversely, building fabs in Korea, which was a popular location for many years, is less favorable.

This season of USD strength may not last, but the effects of it will be long lasting. The most interesting implication, I think, is how China and Japan cooperation can be mutually beneficial. There is a need to establish and maintain trust however.

On the topic of trust: Japan has never built a fab in the US, and Toshiba fended off Western Digital when establishing Kioxia to maintain Japanese control. This is another interesting aspect of the globalization dynamic: Cooperate on one level, compete on another. It's important to always keep this dual dynamic in mind when discussing US-China as well.
 
Last edited:
The semiconductor industry is a model of globalization, with a few countries playing key roles. Japan is a source of advanced materials and automation necessary for fabs to operate, and is perhaps the most important part of this structure. The modern fab depends on robots and materials and Japan has a dominant presence.

This dominance is part technical, part trade. And trade is all about currencies.

USD.KRW bottomed on 11/30/20 at 1085 KRW to the USD. USD.KRW surged to 1347 today, 8/30/22. It has been setting new highs every week for years, and it’s now 24% higher than 11/30/20.
USD.JPY bottomed later than KRW, on 12/28/20, at 103. Since then USD.JPY has been on an epic rise to near new highs on 8/30/22, at 138. It has been less of a consistent rise with some corrections along the way to 34% rise since the end of 2020.
USD.CNH (Offshore Renminbi) has maintained the peg to the USD within a 6.9-7.1 range over the same period.

This brings me to Korea. Korean vendors have certain niches in materials and equipment like Japan that can be exploited when the Won falls.

So US or Chinese fabs purchasing Korean or Japanese goods in Won or Yen have a growing advantage that is leading to more fab construction in these locations. A 24-34% cost reduction for top quality goods is irresistible, both economically (it is worth billions over the life of the fab) and, I think, emotionally (FOMO). Conversely, building fabs in Korea, which was a popular location for many years, is less favorable.

This season of USD strength may not last, but the effects of it will be long lasting. The most interesting implication, I think, is how China and Japan cooperation can be mutually beneficial. There is a need to establish and maintain trust however.

On the topic of trust: Japan has never built a fab in the US, and Toshiba fended off Western Digital when establishing Kioxia to maintain Japanese control. This is another interesting aspect of the globalization dynamic: Cooperate on one level, compete on another. It's important to always keep this dual dynamic in mind when discussing US-China as well.

"China and Japan cooperation can be mutually beneficial"

This is probably an unrealistic hope. When CCP and Chairman Xi shot missiles around Taiwan and circled CCP's Navy destroyers around Japan, the path for cooperations was closed for next 10 years, at least.
 
The semiconductor industry is a model of globalization, with a few countries playing key roles. Japan is a source of advanced materials and automation necessary for fabs to operate, and is perhaps the most important part of this structure. The modern fab depends on robots and materials and Japan has a dominant presence.

This dominance is part technical, part trade. And trade is all about currencies.

USD.KRW bottomed on 11/30/20 at 1085 KRW to the USD. USD.KRW surged to 1347 today, 8/30/22. It has been setting new highs every week for years, and it’s now 24% higher than 11/30/20.
USD.JPY bottomed later than KRW, on 12/28/20, at 103. Since then USD.JPY has been on an epic rise to near new highs on 8/30/22, at 138. It has been less of a consistent rise with some corrections along the way to 34% rise since the end of 2020.
USD.CNH (Offshore Renminbi) has maintained the peg to the USD within a 6.9-7.1 range over the same period.

This brings me to Korea. Korean vendors have certain niches in materials and equipment like Japan that can be exploited when the Won falls.

So US or Chinese fabs purchasing Korean or Japanese goods in Won or Yen have a growing advantage that is leading to more fab construction in these locations. A 24-34% cost reduction for top quality goods is irresistible, both economically (it is worth billions over the life of the fab) and, I think, emotionally (FOMO). Conversely, building fabs in Korea, which was a popular location for many years, is less favorable.

This season of USD strength may not last, but the effects of it will be long lasting. The most interesting implication, I think, is how China and Japan cooperation can be mutually beneficial. There is a need to establish and maintain trust however.

On the topic of trust: Japan has never built a fab in the US, and Toshiba fended off Western Digital when establishing Kioxia to maintain Japanese control. This is another interesting aspect of the globalization dynamic: Cooperate on one level, compete on another. It's important to always keep this dual dynamic in mind when discussing US-China as well.

The reason PRC is screwed in the semi race is that all of its inputs are foreign. Aside from a few own fabs, China has no: microelectronics grade semi boule makers, no final blank wafer makers, no own semi chemistry, no other consumables, no materials handling, no gas/air/water treatment, not own production grade mask shops, facility specialists, fab automation specialists, semi equipment, semi equipment components, and most importantly no know-how holders for how to run fabs as a commercial operation.

Every times optics in a stepper in the mainland fab needs maintenance, it has to fly to Zeiss in Switzerland.

The West/Taiwan/Korea/Japan can shut down mainland's semi industry with a flick of a finger, but not the other way. Most importantly, mainland fabs are hours away from shutdown once the few poached Taiwanese execs making them tick will quit.
 
Last edited:
Paul, TSMC operates a 16nm (finfet) foundry in Nanjing, China. Their 16nm is a very good process and I am told that their yield is quite good. I believe bare wafers can be manufactured here: Changzhou Youze Technology. I haven't checked the other items on your list. There are more knowledgeable people on this forum than me on the manufacturing side. How accurate is Paul's list?
 
Paul, TSMC operates a 16nm (finfet) foundry in Nanjing, China. Their 16nm is a very good process and I am told that their yield is quite good. I believe bare wafers can be manufactured here: Changzhou Youze Technology. I haven't checked the other items on your list. There are more knowledgeable people on this forum than me on the manufacturing side. How accurate is Paul's list?

I knew there are solar wafer makers, and ones who do 200mm, and 300mm for manufacturing of discrete transistors, and such.

IC grade, CMP polished, low defect rate wafers for Nanjing factory, and other mainland fabs are imported as far as I know.
 
I suppose not being able to get more 193nm and 13.5nm lithography machines will be the most limiting factor?
 
I suppose not being able to get more 193nm and 13.5nm lithography machines will be the most limiting factor?

No, the fact is that Taiwan/Japan/Korea each control killswitches for the industry would be much, much bigger hard limits.

Japan — resist related chemistry know-how, wafer handling, amhs, masks
Korea — thin film domination, cleaning
Taiwan — ultralow particle/outgas plastic parts, ultrapure etchants, hyperpure solvents, consumables, test tech, management expertise
 
The reason PRC is screwed in the semi race is that all of its inputs are foreign. Aside from a few own fabs, China has no: microelectronics grade semi boule makers, no final blank wafer makers, no own semi chemistry, no other consumables, no materials handling, no gas/air/water treatment, not own production grade mask shops, facility specialists, fab automation specialists, semi equipment, semi equipment components, and most importantly no know-how holders for how to run fabs as a commercial operation.

Perhaps fabs outside of China don't use Chinese suppliers for these kind of materials,but fabs inside China certainly do. Like most Americans never heard of Xiaomi phone because of it's limited presence in the US,so they just assume that is the case everywhere else in the world.

It's a matter of choice,if you get info from the likes of TSMC NanJing fab of Samsung Xian fab,they may tell you that they don't use any Chinese equipments and materials. But if you get info from the likes SMIC or Huahong,they use tons of domestic equipments and materials.

Every times optics in a stepper in the mainland fab needs maintenance, it has to fly to Zeiss in Switzerland.

Yes,as someone said above, lithography machines is the most limiting factor.

The West/Taiwan/Korea/Japan can shut down mainland's semi industry with a flick of a finger, but not the other way.

You sure about that?There is something vital in semi industry,it's called rare earth. You know how long it takes to find and open new rare earth mine?Decades

The US is working on reducing reliance on China for critical materials,but it won't be easy or quick. So before shut down mainland's semi industry,better make sure that problem is solved
 
Last edited:
Perhaps fabs outside of China don't use Chinese suppliers for these kind of materials,but fabs inside China certainly do. Like most Americans never heard of Xiaomi phone because of it's limited presence in the US,so they just assume that is the case everywhere else in the world.

It's a matter of choice,if you get info from the likes of TSMC NanJing fab of Samsung Xian fab,they may tell you that they don't use any Chinese equipments and materials. But if you get info from the likes SMIC or Huahong,they use tons of domestic equipments and materials.



Yes,as someone said above, lithography machines is the most limiting factor.



You sure about that?There is something vital in semi industry,it's called rare earth. You know how long it takes to find and open new rare earth mine?Decades

The US is working on reducing reliance on China for critical materials,but it won't be easy or quick. So before shut down mainland's semi industry,better make sure that problem is solved

Rare earths are not significant. You lose magnets, polish for CMP, and phosphors, nothing semi related. Ceria will later price correct, as other industries will switch to something cheaper.

CanSemi, and other smaller fabs all bought turnkey solutions from smaller Taiwanese fabs. They are purely a business operation to make money on discrete ICs, and basic MCUs. They don't need experimentation.

SMIC is a whole other story, it being a prestige project with a lot of Shanghai government funding.

Naura, and other fledgling mainland equipment makers all survive thanks to government sponsored fabs being forced to keep them afloat. They are both more expensive (as much as double the price of American/Japanese tools,) and service level/support is said to be terrible.
 
Last edited:
They are both more expensive (as much as double the price of American/Japanese tools,)
THIS IS THE POINT OF THIS THREAD!
It is much cheaper to buy Japanese (when paying in Renminbi)! What an opportunity for a savvy Chinese procurement professional!
The crash in the Yen comes at just the right time to capitalize and save $$$$
 
Also in that article, "Back in July, we reported on an ejection seat problem that grounded more than 100 F-35s, among several other airframes."

Drones and missiles don't need ejection seats.
Or the most ridiculous helmet ever conceived:


Get a different haircut, and it may not fit right. Technology run amuck.

 
Back
Top